Lockdown speculations: Genomics Error Correction.
By Tommaso Demarie, CEO at Entropica Labs.
During these months of lockdown and social distancing, I have taken the habit to go for long walks at dusk. There is an ancient pleasure in solitary contemplation, and every day I take solace from the views of the distant city bathed by the glowing sunset light.
While walking, I occasionally listen to scientific or political podcasts — good ideas always make for excellent companions! On one such occasion, thanks to Azeem Azhar (Exponential View) I familiarised with David Sinclair (Harvard Medical School) and his theory of ageing. Here you can find the podcast episode.
David Sinclair’s work tries to answer the question, Why do we age and why we might not have to. Importantly, Sinclair attempts to answer this via an information theory approach. He proposes that if you boil ageing down to an equation, you end up with information theory and change of entropy in the system. As a long-time practitioner in quantum information theory, this proposition immediately caught my attention.
In this post, I will attempt to explain the critical concepts of Sinclair’s theory, proceed to point them to similar ideas in quantum information, and close by introducing the concept of Genomics Error Correction (GEC). I am here indulging in the speculation of this being a novel concept while claiming that it might become an essential ingredient in any future treatment against the symptoms of ageing.
David Sinclair’s Information Theory of Aging
Information theory meets genomics.
The starting point of the theory, as I understand it, is that epigenetic noise, i.e. information loss, is the primary cause of ageing.
When a complex biological organism is born, it contains an abundance of information. More specifically, there are two types of information, the genome and the epigenome. The genome is the sum of the organism’s DNA, which contains all of the necessary information to build, grow and maintain that organism. The epigenome is the regulation of the genome, also known as gene regulation. Each cell in a body contains the same DNA, but cells can have vastly different functions or identities. The identity of a cell is determined by what genes are switched on or off. The epigenome regulates this choice. The epigenome is essential for development (think of the initial fast differentiation process of embryonic growth) and, Sinclair claims, for long life.
Azhar and Sinclair employ a compelling analogy during the podcast. They say the epigenome is the software (the musical score), and the genome is the hardware upon the score is played (like a piano). During ageing, the pianist loses the ability to perform. For instance, their eyesight might be deteriorating, and the music will not sound pleasant anymore. By the time you are eighty, this has become a cacophony of sounds. What is interesting, Sinclair points out to activate the analogy, is that you cannot replace the piano (the genome) since it is unbuilt. But you could put spectacles on the pianist or maybe bring a new pianist entirely!
if you boil ageing down to an equation, you end up with information theory and change of entropy in the system.
Let us look more at biology. The DNA contains a lot of information (about 6 billion bits in a human DNA), but not all of it is used all the time. DNA is highly wrapped, and most portions remain “hidden” and inaccessible, which means the corresponding genes cannot be read by the cell. These genes are off. During development, some parts unreel so those specific genes can be turned on, which leads to the specialisation of cells. As such, genes switched on and off in the correct order, determine the proper development of an organism.
During ageing, more and more wrapped portions of the DNA start to unravel when they should not, and cells begin to lose their identity. Loss of information breaks the order of the organism. Noise breaks the identity of cells.
A broken molecule of DNA is one of the most devastating error happening in a cell. Intriguingly, these breaks occur all the time, at a rate of trillions per day. The cell has inbuilt, or natural, error correction mechanisms and rapidly moves proteins around to fix these problems. However, the main tasks of these proteins are not fixing broken strands of DNA. Thus, error correction distracts proteins from their functions, which include genes regulation (epigenome work!). So, while there is an active genome repair mechanism inside the cell, Sinclair points out that it comes at the expenses of genes regulation. The regulation becomes neglected, epigenetic noise accumulates, and the cell stops functioning well, causing the symptoms of ageing.
Another way to express this concept is that noise causes epigenetic loss of information about what genes should be activated or not (abnormal gene expression). As a consequence, the cell malfunctions and loses its identity. As the number of these malfunctioning cells increases during a lifetime, the entire organism “malfunctions” more and more, which we perceive as ageing, Sinclair says. Once the number of abnormal cells surpasses a certain threshold, the entire organism fails, and we prosaically call this episode death.
To summarise, in this theory, the loss of information, the inability to read the genome correctly is what drives the causes of ageing. And the causes of ageing drive the disease of ageing, which is ultimately treated in the hospitals.
Explaining a fundamental process like ageing in terms of information theory helps to raise a few natural questions. Can one intervene early and reset the cell? Could we reboot the software and potentially undo ageing? In other words, can we develop a theory of active Genomics Error Correction to keep the error rates below a certain threshold such that cells do not lose their identity and the causes of ageing do not manifest (or manifest more slowly)?
Quantum error correction
On making a conceptual connection between biological ageing and quantum computing.
A quantum computer is a programmable device that processes quantum information. One way to picture the process is to imagine it as a sequence of steps in time, each one corresponding to a discrete set of operations, also known as gates, performed on the quantum systems (the piano) according to a quantum algorithm (the music score, if you like).
When the algorithm is well designed, this process takes us from a specific input to the desired output. Unfortunately, with our current technology and understanding, this holds only in theory. In practice, the quantum systems that contain the information are very fragile. When we act on them, they quickly break down (decohere) eventually resulting in the loss of all information if the computation requires too many time steps.
Hence, to build reliable quantum computers, it is very likely we will require strategies to correct for errors that emerge while operating on quantum systems. Quantum Error Correction (QEC) is the theoretical framework of codes and techniques used to correct the errors that plague a quantum system during its evolution. These methods make use of a vast range of concepts (like redundancy, detection, concatenation) to ensure that a computation is carried forward as the theoretical algorithm, its instructions, demands.
Decoherence intuitively resembles ageing in the sense of information loss. Without QEC, the interaction with the environment causes the computation to reach an irreversible state, similar to a biological organism that interacts with the environment and eventually reaches death (as far as we know, also somewhat irreversible).
If I were to make a feeble parallel to the analogy of the pianist above, one could understand QEC as a process where at each time step, we replace the pianist. At the same time, the music score (the algorithm) and the piano (the quantum systems in the processor) remain the same. Today, QEC is inseparable from quantum computing and an essential aspect of quantum information theory.
From QEC to GEC
The thrill of taking things too far!
Let me speculate now. Assume this theory of information offers a practical approach to understanding and acting on the biology of ageing. Assume we reliably identify the pathways (and proteins) responsible for epigenome information loss. Assume it is possible to actively restore the information loss at the cellular level such that cells do not lose their functional identity (Sinclair has some suggestions on how this could be achieved).
If you are comfortable with all these assumptions, it seems reasonable to term this framework Genomics Error Correction (GEC) and imagine that conceptually it is of kin to QEC. Worry not my physicist friends, I am not implying that quantum effects play any role in genomics or GEC. I am instead fascinated by the prospect of borrowing a robust and fundamental approach (QEC) and suggest that a similar method (GEC) could become critical in our future understanding of treating conditions due to ageing.
I can imagine in a not-so-distant future, doctors administering molecules (maybe proteins directly — protein therapeutics) to stimulate or control the pathways responsible for gene regulation, while natural QEC takes care of DNA breaking. Such a scenario could become a routine in the very human fight against the much-feared symptoms of ageing.
Since we are speculating, why not going one step further down the rabbit hole. In quantum computing, one marvellous result is the Fault Tolerance Theorem, or Threshold Theorem, which states that — due to the mathematical structure of quantum mechanics, there exists a regime where you can correct for errors faster than they accumulate.
I then feel compelled to ask. Could there exist a genomics analogue of the Threshold Theorem, which states that if the rate of creation of errors — as intended by the information theory of ageing — is small enough, it is possible to correct them actively and hence stop an organism from deteriorating?
While walking at dusk, enjoying the cooling evening breeze, it is tantalising to believe that the answer might well be yes.
Warnings: This is a notional Sunday post, not a peer-reviewed article and as such, its claims are likely to be shaky. To the best of my knowledge, error correction in genomics traditionally refers to the techniques used in sequencing to correct for errors while reading the genome. Here we have proposed a different meaning for the expression.
I believe it is powerful to have molecular biologists reasoning in terms of information theory, a modus operandi that has been very successful for quantum physics. I hope I can offer a minuscule contribution by suggesting that it might be beneficial to add concepts from error correction to the picture. Happy for experts to prove I am, in order, naive, simplistic, or just plain wrong. Mistakes, abuses of concepts, misunderstanding and ideas taken too far are mine and mine only.
For the interested readers, David Sinclair has written an excellent book about his theory, titled “Lifespan: Why We Age — and Why We Don’t Have To”.